A hierarchical framework for understanding multiple- and single-action representation

• Applicant: Dr. Tim Raettig
• Subject Area: General, Cognitive and Mathematical Psychology
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 573208198

Previous research in the action-control domain has either investigated multiple-action control or single-action preparation, but not how both interact. Theoretical frameworks that can explain the occurrence of dual-action costs and benefits are, thus far, disconnected from approaches towards explaining single-action planning and modification. As a result, a comprehensive account of multiple- and single-action control and, in particular, action representation is currently lacking. As a tentative solution, we have developed an integrative theoretical account of hierarchical action representation (the Task-Tier-Type-Token Framework, in the following: "four-T F", 40F) which we want to further refine and test in the empirical studies outlined in this proposal. In a nutshell, the 40F holds that 1) actions are cognitively represented as bundles of distinctive features (e.g., “effector”, “laterality”) and 2) these features are hierarchically structured (e.g., effector is specified before laterality). Procedurally, it is assumed that when an action plan is formulated, the first decision – on the tier level – is about the number of individual actions that have to be prepared (i.e., none, one, two, and so on). Essentially, the idea is that space in working memory is reserved (and possibly pre-structured) in terms of different “slots” for motor programs that are later specified on lower representational levels. Tier-level effects will be investigated in Work Package 3. Once tier-level selection is completed, the corresponding protoprogram slots are further specified on the type level by deciding on particular effectors and patterns of muscular innervation to be used. Type-level effects will be investigated in Work Package 2. Finally, on the token level, motor program parameterization is completed by setting rank-and-file features such as laterality, extent, and so on. Token-level effects will be investigated in Work Package 1. While current research on action control is mostly devoted to higher-level mechanisms (e.g., switching between fully specified task sets or multi-tasking), the approach presented here addresses the more fundamental, fine-grained inner workings of action representation and strives to ultimately arrive at a solid, well-specified basis that can then be integrated with the aforementioned overarching accounts. Theoretically, we are building on the well-established protoprogram concept discussed by David Rosenbaum as well as modern feature-based approaches towards action representation, integrating these with ideas regarding lower-level representational hierarchies. Empirically, we will extend an experimental paradigm established over the course of two previous projects investigating single- and dual-action precuing, taking data from related studies on multiple-action control in other settings into account.

• DFG Programme: Research Grants
• Co-Investigator: Professor Dr. Lynn Huestegge